The determinants of household disaster preparedness behavior in Bangladesh
Recent literatures have widely portrayed natural disaster impacts on welfare and livelihoods. This literature largely focuses on loss of physical capital (e.g. asset, livestock, and crop), household income earning opportunities, coping strategies (including migration) and health outcomes. The aspects that are extremely less understood are the role of human capital (i.e. direct and indirect effects of education) and its impact on disaster risk reduction policies through adoption of disaster preparedness measures in lower and middle-income countries with high climatic risks. This study aims to add to the growing “climate-development” literature by investigating the possible determinants of household disaster preparedness behavior particularly focusing on knowledge and perception and prior damage and employment channels of disaster experience utilizing the BBS 2015 Impact of Climate Change on Human Life (ICCHL) survey which is a unique and comprehensive large scale big data in Bangladesh.
The supply side focus of the government policies e.g. cyclone shelter, embankment construction etc. has been appreciated both nationally and internationally specially with respect to mortality reduction and minimizing loss and damages. However, there are extremely limited discussions on the effectiveness of the demand-side policies of the GOB i.e. disaster preparedness demonstrating their impacts on disaster risk reduction in the short- to medium-longer term at the household level. Hence, there exists a strong need to scrutinize the existing short- to medium-longer term disaster risk reduction policies of the Bangladesh government for efficient domestic and international investments for successful implementation of the Climate Action Plan i.e. SDG 13.
In the light of our strategic focus on the longer-term solutions to reduce climate and disaster risk at the micro/household level; our research questions have been established to look at the effectiveness of the key policy interventions of the Bangladesh government in DRR through knowledge, perception, education, social capital, infrastructure and household loss mitigation. We looked at three (3) policy interventions of the government; namely, existing early warning system, public awareness for preparedness and government disaster financial support and attempted to answer two (2) research questions. We first identify the determinants of household disaster preparedness behavior including justifying the role of climate change and disaster knowledge and perception, human and social capital, disaster displacement, remittance, income, wealth and labor market outcomes. We then attempt to look at the effectiveness of the key government policies and responses in mitigating household loss and damage to sustain economic development at the household level. To incorporate the respective policy responses, our methodological approach has been primarily index creation targeting the policy relevant questions’ in various modules of the ICCHL survey and thereby include in our empirical specifications for estimation purpose. For example, we initially develop three (3) preparedness indices i.e. evacuation preparedness index (to clarify the extent of the existing early warning policies influencing household evacuation preparedness measures), preservation preparedness index (to clarify the extent of households’ adoption of preservation preparedness measures due to public awareness adhering to the global 72-hour emergency supplies rule) and infrastructure preparedness index (to clarify the extent of government’s disaster financial support e.g. government loan support influencing households’ infrastructure preparedness measures).
This study provides strong context-specific empirical evidence on the possible determinants of household disaster preparedness behavior utilizing a unique and comprehensive big data in Bangladesh. Our results show that, in Bangladesh, disaster risk perception explains around 1.46% of formal education with nearly 7.39% of female education compared to the mean. Similarly, climate change perception accounts for around 1.51% of formal education whereas education levels of female household heads explains almost 7.68% compared to the mean indicating the overwhelming importance of enhanced disaster education (formal and non-formal) for women in climate and disaster risk reduction policies in Bangladesh. We argue that mainstreaming disaster education across education, climate change and disaster risk reduction policies could significantly enhance responsiveness to our disaster preparedness behavior and thereby increase household resilience.
We also found that household responses to adoption of numerous forms of preparedness measures ranges between 2%–3.6% due to increase in remittances while ranging between 1.4%–2.4% increase due to percentage increase in residing in better housing (i.e. resilient infrastructure). Access to safe drinking water, electricity and sanitation also influences preparedness behavior significantly. Our findings show that social capital is a robust determinant of adoption of preparedness measures. We identify disaster displacement as an important determinant of disaster preparedness behavior of households. Our positive and significant result of disaster displacement variable indicates that displaced people are found to be better prepared due to their past disaster experience and actively respond to government interventions and policies. Our results represent that wealth and salaried income are positively and significantly associated with disaster preparedness compared to net per capita income and daily wages. Per capita income (net) is found to have a negative relationship (but insignificant) along with daily wages which is also not found to be strongly correlated with disaster preparedness behavior.
Our contribution to the “climate-development” literature is three-fold: first, we argue that in the absence of a globally agreed loss and damage framework, this study conceptualizes disaster preparedness as a risk reduction pathway towards sustainable development. Second, we provide strong empirical evidence to what extent disaster preparedness could reduce household loss and damage arising from 12 types of natural disasters among 143,980 households across 64 disaster affected regions (including agro-ecological zones) of Bangladesh using big data for the first time in this study. Third, we strongly argue that integration of development and disaster risk reduction policies could further reduce the amount of losses arising from climate-induced natural disasters implying integrated impacts across various SDG targets at the micro-household level. Our findings suggest that disaster preparedness is almost 76% effective in mitigating net income loss (per capita) and nearly 81% effective in mitigating salaried income loss (annual) arise via unemployment channel (i.e. loss of employment days due to climate disasters) at the household level.
To conclude; despite widespread policy successes at the local and international contexts, it has been evident that non-adherence to some of the government policies still remains which translates to their ineffectiveness at various contexts. For example, vulnerable people are often found reluctant to leave their homes and assets due to the lack of security despite early warning or they often look for high lands nearby their houses or take shelter on the embankments rather than flood shelters. A recent high-level consultation meeting delineated that public awareness interventions might not be fully effective without high perception and knowledge level regarding climate change and disaster risk impacts in the short- to medium-longer term. This again reiterates the fact that households’ responses (demand-side) and government’s actions (supply-side) complement one another in most cases and efficacy of household responses could deliberately depend on other public interventions and responses as well. In some cases, effectiveness of the demand-side policies in terms of the uptake of preparedness measures might depend on the behavioral responses of the households. We assume that uptake of preparedness measures due to policy intervention of early warning information might depend on the timing of disaster occurrences i.e. deliberation of the “what to do list” in formal education (via textbooks) might not lead to full effectiveness of these policies in many cases; rather, informal education and short-term skill based designated education programs might address several of these issues including enhanced resilience gained through knowledge and perception. We therefore recommend short-term and disaster-specific “72 hour early warning-based preparedness education program” and/or “3-5 day flood forecasting model-based preparedness education program” as potential solutions that requires further research.
STUDY DIRECTOR: DR. AZREEN KARIM, RESEARCH FELLOW, BIDS.